A systematic meta-review of organizational diversity and inclusion interventions and their associated outcomes

There are many frameworks for fostering diverse and inclusive workplaces, and extensive research has highlighted the systemic and multilevel aspects of D&I practices (e.g. Mor Barak, 2000; Shore et al., 2018). While many theoretical frameworks have painted a comprehensive picture of the components of diversity management, they translate less well in the practical context where organizations want to know what factors they can change, which is critical for practical guidance. One particular model that emphasizes practical relevance is Kulik’s (2014) five components of organizational diversity management systems, which reveals the gap between academic research and the informational needs of practitioners.

Kulik (2014) contends that three diversity system components are related to clear actions that leaders and managers can take: paradigms (an organization’s values, beliefs, and norms about diversity management), policies (organizational goals or objectives for managing diversity), programs (formal diversity management activities). The other two diversity system components—practices (the implementation of diversity programs by lower-level staff) and diversity climate (shared organizational perceptions of diversity paradigms, policies and programs)—are largely measured through micro-level employee perceptions and experiences. While HR research has concentrated heavily on the latter two components, Kulik’s (2014) model directs us to move beyond “measuring things that may be important but ‘that managers cannot manipulate’” (Kulik, 2014, p. 134).

We thus turn our attention to actions that organizations can take, some of which have already been standardized and categorized within existing frameworks in the diversity management literature. For example, Shen et al.’ (2009) diversity management framework is based on traditional categories of HR management practices, including recruitment and selection (e.g. affirmative action programs), training and development (e.g. diversity training), performance appraisal (e.g. objective performance-focused appraisal systems), and pay (e.g. individual-based remuneration systems). Other theoretical frameworks are based on the mechanisms through which activities are thought to achieve diversity, such as non-discrimination practices (e.g. merit-based or de-biased decision making, diversity training; Leslie, 2019). A common approach is categorizing “bundles” of policies and practices—interrelated and synergistic practices thought to bring about greater productivity when combined (MacDuffie, 1995; see also Boehm et al., 2021; Richard et al., 2013). Importantly, these frameworks—while conceptually useful—have limited practical applicability where they have stopped short of showing empirical links to D&I consequences. Shen et al. (2009, p. 237) noted that “there has been no empirically proven association of HRM diversity practices with increases in diversity and improved organizational performance”.

While there is no shortage of popular articles on “what works” with respect to organizational D&I activities, HR managers and organizational leaders wanting to draw on evidence-based practice may face several barriers. First, while many interventions have been put forward—diversity training, standardized tests, metrics, mentoring, and appointing diversity managers and taskforces (Bohnet, 2016; Dobbin and Kalev, 2015, 2016; Kalev et al., 2006; Kulik and Roberson, 2008; Pedulla, 2020; Roberson et al., 2020)—the maturity and rigor of the research is mixed. There has been little synthesis of knowledge to help time- and resource-limited practitioners weigh up the likely effectiveness and quality of evidence supporting the range of actions that they can take.

Second, the focus on theoretical and conceptual work has meant that less is known about empirical outcomes, with research dominated by diversity-specific outcomes (e.g. representation of women at recruitment; Dobbin and Kalev, 2016). As such, practitioners risk investing in initiatives under the broad umbrella of D&I with no clear goals and without knowing what positive (or negative) outcomes to expect (Nishii et al., 2018).

Finally, as Ayoko and Fujimoto (2023) note, there is “inequity” in how research has engaged with the different dimensions of diversity. Within the HR management, research over the last 60 years has largely concentrated on gender and racial discrimination (Triana et al., 2021). Procknow and Rocco (2016) reported that disability—particularly cognitive and mental—is often omitted from definitions and discussions around diversity. Research on workplace accommodations for age and disability inclusion largely sits within the domain of occupational rehabilitation and public health (e.g. Nevala et al., 2015). In reality, organizations often wish to improve D&I efforts across the spectrum of diversity dimensions—usually gender, race, age, and disability, but as many as 27 different dimensions based on corporate diversity statements (Point and Singh, 2003). The lack of cross-pollination between disciplines has meant that there is a shallow pool of research for comparisons on what interventions work best (Kulik, 2014). Pitts (2006) argued that the “fragmentation of diversity research is arguably its biggest problem” (p. 262). In sum, practitioners wishing to quickly access what actions are effective may find research locked in disciplinary siloes, with publications “driven by theoretical and discipline-specific concerns” (Kougiannou and Ridgway, 2022, p. 472).

This study aims to address the gaps identified in the previous section by providing a unifying summary of D&I interventions extending beyond the HR management literature and across multiple disciplines. It asks the following questions: (1) What types of organizational D&I interventions have been examined in the academic literature and how effective are they? (2) What outcomes are empirically linked to these interventions? (3) What diversity dimensions are targeted by these interventions?

Given the practical focus of this research, the broad and heterogeneous scale of the literature, and existing pockets of established research (some which have already produced several systematic reviews), a meta-review was selected (Hennessy et al., 2019). Researchers, policy-makers, and practitioners seeking to translate scientific findings for applied contexts encounter significant challenges and practical difficulties in synthesizing the large volume of research spanning multiple disciplines (Caird et al., 2015; Khangura et al., 2012). In light of these barriers to feasibility, the meta-review approach is designed to overcome limitations of traditional systematic reviews by providing a timely, succinct, and methodical summary for real-world decision making. In addition, meta-reviews “can have tremendous influence on research, practice, and policy; indeed, if conducted appropriately, synthesis literature is considered the strongest level of evidence, with meta-reviews atop the evidence pyramid” (Hennessy et al., 2019, p. 354).

In the following, this meta-review aggregates and synthesizes evidence from diverse bodies of literature on the impact, outcomes, and quality of evidence to guide organizations taking action for workplace D&I. We begin by outlining the methodology, followed by a presentation of findings categorized by intervention type, and conclude with a discussion of implications for both research and practice.

In this section we present our method based on the main stages of a meta-review (Biondi-Zoccai, 2016), starting with a search strategy followed by quality appraisal of articles. This is followed by a description of our data extraction and coding of information. For full methodological details, see the Online Supplementary Material.

A protocol consisting of search terms and inclusion/exclusion criteria for this review was developed a priori by the authors in consultation with an academic reference librarian. A comprehensive search was undertaken of five databases (PsycINFO, Business Source Complete, Web of Science, Cochrane Library, and Campbell Collaboration Library) in January 2019 and updated in November 2021, for all articles published after 1 January 2009. Inclusion criteria consisted of population and setting (employees in any organization and in any industry in developed countries that are practically comparable to Australia or similar contexts), study designs included in reviews (interventional or observational), and intervention type (activities designed with the goal of D&I).

Title and abstract screening and full-text review were carried out independently by the first two authors (KZ, WH) based on the inclusion/exclusion criteria in January 2019, with discrepancies resolved through discussion, and later by one author (KZ) in November 2021.

Quality appraisal was conducted in duplicate by two authors (KZ, WH) using AMSTAR (A MeaSurement Tool to Assess systematic Reviews; Shea et al., 2007), a critical appraisal instrument for evaluating the methodological rigor and quality of systematic reviews and meta-analyses. Not all included articles were systematic reviews and eligible for AMSTAR, so articles were also divided into a two-tiered system based on the hierarchy of evidence (Murad et al., 2016): (1) systematic reviews and meta-analyses, and (2) narrative reviews and scoping reviews.

A number of fields were extracted and articles were coded according to the specific intervention for improving D&I. Where multiple interventions were mentioned, these were coded according to the narrowest distinct intervention reported and care was taken to separate the activities so that they were mutually exclusive. Interventions were then grouped into 12 broader categories for ease of interpretation (see Figure 1).

In addition, outcomes associated with each intervention were also recorded. For example, extending the working lives of older employees, increasing the share of women on boards, and career advancement of staff from minority backgrounds were coded to the outcome Diversity in recruitment, retention, or progression. Each intervention–outcome pairing was also coded according to whether the activity was reported by the review authors to have (1) evidence of being effective (or the authors recommended the intervention based on the evidence), (2) evidence of being weakly effective, (3) evidence of mixed positive and negative effects, (4) evidence of no effect, (5) an unknown effect, a lack of evidence, or the authors did not specify the direction of the effect.

In this section, we first present a summary of the search results, followed by findings organized by the three research questions, and for each of the D&I intervention categories.

Figure 2 illustrates the search process using a PRISMA (Preferred Reporting Items for Systematic reviews and Meta-Analyses; Page et al., 2021) diagram. The PRISMA diagram depicts a transparent summary of the search process and is required to meet established standards for reporting in systematic reviews (Page et al., 2021). The initial search yielded a total of 6,769 citations, after removal of duplicates. Full-text review produced a final list of 37 accepted articles, comprising 4 meta-analyses, 17 systematic reviews, 7 scoping reviews, and 9 narrative reviews (see Table S1 in the Online Supplementary Material). Articles varied in quality, with only one meeting all eligible AMSTAR criteria (Gensby et al., 2014) and five meeting three or fewer criteria. Articles came from a range of fields including HR, organizational behavior, applied psychology, vocational rehabilitation, occupational health, and public health.

From the 37 articles, we identified a total of 60 specific practices, policies, programs, or initiatives for organizational D&I, which were grouped into 12 categories. We also identified 22 outcomes which were linked to one or more of these interventions (see Table S2 in the Online Supplementary Material).

Figure 3 presents the number of articles that reported on D&I interventions across the 12 categories, broken down by the type of evidence presented. Diversity training and education, Policies involving recruitment, leave, and compensation, and Workplace accommodations were the categories with the most articles (12, 12, and 14 respectively). Longer bars in the figure do not necessarily mean that the category was more effective, rather that the interventions were examined more frequently. For example, more than half of the interventions belonging to Policies involving recruitment, leave, and compensation were reported to have mixed effects, evidence of no effect, or a lack of evidence. In contrast, Workplace accommodations were found to be more consistently positive, with 12 of 14 articles reporting them to be effective. The dotted line provides a basic indication of the hierarchy of evidence, showing the percentage of articles within the category that were systematic reviews or meta-analyses. Diversity training and education had relatively higher quality of evidence compared with other categories, with a higher number of articles and more than half of articles being meta-analyses or systematic reviews.

Due to the search criteria, Diversity in recruitment, retention, and progression was by far the most common outcome associated with D&I interventions, and was reported for all categories. Some categories of interventions—Diversity training and education, for example—were linked to a range of outcomes, such as learning and behavior, among others. In contrast, De-biasing of selection procedures was associated with only diversity outcomes.

Age, disability, and gender were the most common specified dimensions of D&I, although ethnicity and race were examined in a number of general D&I articles, particularly those examining diversity training. In contrast, only one article exclusively focused on lesbian, gay, bisexual, and transgender (LGBT) D&I (Webster et al., 2018). Categories such as Workplace accommodations and Health promotion were limited to age- and disability-based diversity. In contrast, De-biasing of selection procedures were widely examined with respect to several dimensions of diversity. Specific findings and evidence quality for D&I intervention categories are reported below.

Audits, reviews, and assessments included mandatory organizational assessments on topics such as the age composition of the workforce (Dickson, 2015). The quality of the evidence for these interventions was average: they tended to be reported as recommendations and were associated with general D&I outcomes.

De-biasing of selection procedures included practices such as blinding gender information in the recruitment process and standardized behavioral interviews (Alonso et al., 2017; Boehm et al., 2021; Isaac et al., 2009; Schloemer-Jarvis et al., 2022; Yang and Konrad, 2011). Most interventions in this category were associated with outcomes promoting diversity in recruitment, retention, and progression, and generalized across a number of diversity dimensions. For example, Isaac et al. (2009) presented a number of effective interventions in their systematic review of approaches for eliminating gender bias during personnel selection, including focusing on job-related competence recruitment, ensuring that women comprised 25% or more of the candidate pool, and committing to job criteria before applicants were reviewed. The quality of evidence associated with this category was higher than others, with more than half of articles consisting of systematic reviews and meta-analyses.

Although diversity training has been criticized for having a theoretically-limited and methodologically-flawed evidence base (Alhejji et al., 2016; Kalinoski et al., 2013), the quality of evidence in this category was relatively higher than at least the other categories in this review. More than half of the articles were meta-analyses (Bezrukova et al., 2016; Kalinoski et al., 2013) and systematic reviews (Alhejji et al., 2016; Isaac et al., 2009; Metinyurt et al., 2021; Phillips et al., 2016; Schloemer-Jarvis et al., 2022), with an average AMSTAR score of 6.

Diversity training was typically linked to three major outcome categories: (1) changes in cognitive outcomes, such as knowledge and awareness, (2) changes in attitudes, emotions, and motivations, and (3) changes in behavioral skills and intentions, and to a lesser extent, changes in behavior (e.g. Alhejji et al., 2016). Beyond these, the evidence for training and education to promote diversity in recruitment, retention, and progression was weak. Three articles reported on a lack of evidence or no effect of diversity training on gender equity in hiring and employment-related outcomes for people with disability (Isaac et al., 2009; Phillips et al., 2016; Schloemer-Jarvis et al., 2022). However, some authors pointed to the importance of education as an enabler of disability inclusion. Cavanagh et al. (2017), McDowell and Fossey (2015), and Nevala et al. (2015) all noted the need for education to overcome employers’ lack of awareness of financial support, workplace accommodations, discrimination legislation, and intervention strategies.

Flexible work arrangements were investigated with respect to improving age (Boehm et al., 2021; Dickson, 2015; Edge et al., 2017; Mahon and Millar, 2014), disability (Nevala et al., 2015), and gender (Gabaldon et al., 2016; Hegewisch and Gornick, 2011) dimensions of diversity. However, some authors cautioned that these practices could backfire where they prevent progression of women to senior positions or increase reluctance among employers to hire women (Gabaldon et al., 2016). Other authors recommended that flexible working arrangements need to be accompanied by greater transparency and visibility to all employees, including how they may affect opportunities for promotion or salary increases (Lane and Flowers, 2015).

Health promotion activities refer to a range of initiatives involving health checks, health counseling, health screening (Söderbacka et al., 2020), occupational health programs (Wilson et al., 2020), and other activities to encourage healthy lifestyles (Boehm et al., 2021; Cloostermans et al., 2015; Edge et al., 2017). These activities have been recommended predominantly as a means of improving the health and work ability of older workers, preventing early retirement, and extending careers for age-based diversity. Although research samples in this category were limited, there was some agreement that they improved work outcomes for older adults (Söderbacka et al., 2020).

Workplace supports refer to supportive actions and relationships with managers and supervisors, along with co-workers (Boehm et al., 2021; Jansen et al., 2021; Nevala et al., 2015; Webster et al., 2018). Webster et al. (2018) and Beehr and McGrath (1992) distinguish between three types of social support: emotional (e.g. showing concern, empathizing), instrumental (e.g. intervening as a bystander, corroborating reports of harassment), and informative (e.g. providing advice on addressing workplace discrimination). There was reasonably strong evidence to link these practices with inclusion-relevant outcomes (e.g. improved work attitudes, reduced psychological strain) for LGBT employees.

Job training, coaching, and vocational counseling provided by employment support specialists were common initiatives for improving age and disability dimensions of diversity by facilitating employment (Boehm et al., 2021; Carlson et al., 2020; Cloostermans et al., 2015; Dickson, 2015; Khalifa et al., 2020; Mahon and Millar, 2014; McDowell and Fossey, 2015; Nevala et al., 2015; Procknow and Rocco, 2016; Schloemer-Jarvis et al., 2022). Lack of job-relevant investment in skills—especially in using technology and the changing demands of the workplace—was a major barrier to continued employment among the aging workforce (Dickson, 2015; Mahon and Millar, 2014). Procknow and Rocco (2016), for example, suggested that personalized training and career planning development can improve employees’ self-determination, self-confidence, and career satisfaction.

Although one of the larger categories in this review, more than half of the 12 articles involving formal organizational policies showed mixed effects, no effect, or a lack of evidence or specificity around known outcomes. Gabaldon et al. (2016) found that while gender quotas improved women’s access to more senior positions, they were also associated with unintended consequences such as an organization exiting into an organizational form that is not regulated by gender quotas. One review found that equal opportunity directives and statements increased organizational attractiveness to female and minority candidates (Yang and Konrad, 2011), although another suggested that they did not ensure gender equity in hiring and salary recommendations (Isaac et al., 2009). Webster et al. (2018) examined the effect of formal policies and practices which were weakly associated with outcomes among LGBT workers. The authors also pointed out the signaling role of formal policies, which can be a visible representation of organizational norms (Webster et al., 2018).

Mentoring refers to formal or informal relationships in an organization where a senior employee provides guidance in the form of knowledge, expertise, and skills to a less-experienced employee. Role models allow for professional identification and emulation among more junior workers from diverse backgrounds. This is especially important given evidence that minority workers report greater barriers at the advancement stage (vs. the career entry) where there is a lack of mentors and personal networks (Kilian et al., 2005).

Role models, mentoring, and sponsorship were recommended across a range of diversity dimensions (e.g. Gabaldon et al., 2016; Kapoor and Solomon, 2011; Procknow and Rocco, 2016). Gabaldon et al. (2016) noted the difficulty in measuring the role of mentoring given the small number of women in leadership and time lags in outcomes. More recently, the practice of reverse mentoring—whereby more experienced workers are in the roles of mentees paired with more junior workers—has shown some promise in educating about the experiences of minorities in biased systems (Chaudhuri et al., 2021).

Two articles recommended strategies involving greater visibility of employees from marginalized or disadvantaged groups. Gabaldon et al. (2016) discussed the importance of visibility of women in leadership positions—such as through online databases—to overcome barriers such as identification with gender role expectations. Visibility can also be implemented at the recruitment stage, such as pictorial displays of minority groups in advertising and involving demographically-similar recruiters (Yang and Konrad, 2011).

Workplace accommodations consist of modifying work schedules and work environments, restructuring job duties, adjusting supervisory methods, and acquiring specialized equipment (Job Accommodation Network, 2022). It was the largest category of interventions in this review, consisting of 14 articles, and the quality was among the highest (average AMSTAR score of 6.2). Nevala et al. (2015) found moderate evidence of specific workplace accommodations (e.g. vocational counseling, change to work schedules, education in self-advocacy) for promoting employment among workers with a physical disability. Among employees with a mental illness, however, McDowell and Fossey (2015) found limited evidence that workplace accommodations resulted in longer tenure.

This meta-review spans across disciplines and diversity dimensions to examine a range of D&I interventions and their associated outcomes. We identified 37 reviews published over 13 years, which included 12 categories of practices, policies, programs, initiatives, or other interventions. These interventions were mapped to 22 distinct outcomes and showed varying effectiveness and evidence quality (see Table S2 in the Online Supplementary Material).

We bring these elements together in a practical framework for guiding the selection of organizational D&I interventions (Figure 4). While this framework shares some interventions with existing diversity management frameworks (e.g. Shen et al., 2009), it extends them by taking a multi-disciplinary approach that is developed bottom-up based on empirical links with D&I outcomes, producing a wider set of intervention categories and evidence-driven outcomes. In this framework, interventions are grouped into high-level categories and linked to empirical outcomes. These links are evaluated by a number of criteria: the volume and quality of evidence, direction and size of effects, and presence of mixed effects.

By overcoming conventional boundaries in discipline- and dimension-specific research, this meta-review has revealed varying levels of maturity in understanding workplace interventions for D&I. It also highlighted a number of influential studies in these respective areas. Kalinoski et al.’s (2013) meta-analysis of 65 studies examined the affective, cognitive, and skill-based outcomes of diversity training, along with their boundary conditions and moderators. Likewise, Bezrukova et al.’s (2016) broader meta-analysis analyzed 260 studies against a theoretical framework, revealing positive effects of diversity training particularly for reactions to training and learning, and under certain conditions or delivered in specific formats. Beyond diversity training, Hegewisch and Gornick’s (2011) narrative review is an example of the unintended consequences of D&I interventions, which highlighted inequalities in employment hours and pay between mothers and fathers, and the impact of long paid maternity leave policies that may reinforce this. Amid growing interest and research on sexual orientation diversity management, Webster et al.’s (2018) meta-analysis provided an early useful framework of interventions in the form of workplace contextual supports (e.g. formal policies) and outcomes (e.g. disclosure) for LGBT employees. More broadly, Yang and Konrad’s (2011) narrative review provides a higher-level summary of the antecedents and impacts of diversity management practices by drawing on institutional and resource-based theories.

From this broad evidence base, this meta-review has highlighted several key findings. No single category of intervention outperformed all others, but there were several tradeoffs across the quality of evidence, the range of D&I outcomes, the effectiveness of interventions, and the generalizability of results across diversity dimensions. While the volume of evidence was weighted towards some interventions—for example, workplace accommodations and diversity training and education—these interventions showed shortcomings in other respects. Research on diversity training and education, for example, tended to focus on circumscribed awareness, knowledge, and learning outcomes (e.g. learning about biases) rather than organizational change. The literature on recruitment, leave, and compensation policies showed mixed effects on D&I outcomes. Practical approaches such as workplace accommodations and job training that provided workers with resources and skills were more consistently associated with positive outcomes, but concentrated on age and disability diversity only.

Managers and leaders face considerable challenges in carrying out D&I initiatives, where the need for meaningful organizational change may be met with limited resources, as well as the risk of tokenism, inauthenticity, and backlash. This meta-review has identified directions where resources can be spent efficiently on actions that have the most effect. One major theme from this meta-review was the specificity of links between D&I interventions and over 20 outcomes ranging from learning, career satisfaction, health, and reduction of psychological strain. For example, diversity training was associated with proximal reactions and learning outcomes (Bezrukova et al., 2016; Kulik and Roberson, 2008) but there was little evidence that it translated to widespread diversity in recruitment (see also Kalev et al., 2006).

These findings explain why many organizations fall short of widespread change, as they are at odds with how D&I are often treated in practice. Operationally, diversity and inclusion are often used interchangeably and short-hand as a common banner for any policies or practices relating to aspects of D&I (Jonsen et al., 2021). Many D&I activities, such as education programs, are piecemeal and “band-aid” type actions without a clear connection to an overarching diversity management framework. The current findings help organizations investing in diversity training (or other interventions) set realistic expectations about what they can achieve, and encourage a suite of interventions targeting multiple aspects of diversity (see also Kulik, 2014; Stephens et al., 2020). The findings also help practitioners anticipate where unintended consequences and mixed effects may occur (for a theoretical review, see Leslie, 2019).

Finally, the findings provide insights on the conditions for effective interventions through moderating factors and enabling environments. In the case of diversity training, instructional method (e.g. active, such as roleplay vs. passive, such as reading) and the length of training (e.g. more than 4 h) were critical (Bezrukova et al., 2016; Kalinoski et al., 2013). For other intervention categories such as policies, cultural, political, and economic context were proposed key moderators (Alonso et al., 2017; Hegewisch and Gornick, 2011; Schloemer-Jarvis et al., 2022).

In assessing the state of the evidence on D&I interventions, this meta-review has revealed some gaps in the literature. Some categories (e.g. diversity training and education) were examined with meta-analyses over a hundred studies, and others had virtually no reviews despite compelling arguments from primary research, which would benefit from future tests of interventions (e.g. accountability; Dobbin and Kalev, 2015; Richard et al., 2013).

There is also a need to consider LGBT inclusion in particular (Webster et al., 2018), which has unique considerations (e.g. barriers relating to disclosure) that are not generalizable. Within the disability dimension of diversity, there were few studies investigating accommodations for “invisible” disabilities (i.e. cognitive and psychiatric disabilities; Prince, 2017; Sundar, 2017). It is also encouraging that a recent review has identified a range of activities—accommodations during recruitment, modified onboarding processes, job design practices, flexible working, training, and employee resource groups—for promoting neurodiversity (Hennekam and Follmer, 2024).

Finally, organizational practices have traditionally focused on diversity and, despite shifting interest to inclusion (Mor Barak et al., 2016), reviews have rarely examined inclusion-relevant outcomes, such as perceptions of belonging, uniqueness, authenticity, participation, and access to resources (Jansen et al., 2014; Mor Barak and Cherin, 1998; Roberson, 2006).

We used a rapid meta-review methodology (Khangura et al., 2012) because it is ideal for synthesizing broad bodies of literature in an efficient and methodical fashion. In covering the breadth of research, the tradeoff is depth on the myriad practices, outcomes, and diversity dimensions associated with organizational D&I. It is reassuring, however, that rapid reviews generally produce similar conclusions to systematic reviews and are a sufficiently reliable source for decision-making (Watt et al., 2008).

Nevertheless, the broad search strategy should be kept in mind when interpreting the findings. Effective D&I strategies that do not lend well to interventional study designs are at risk of being overlooked. Some promising approaches that are of practical significance are missing from this review, such as independent ombudspersons (Dobbin and Kalev, 2019), diversity task forces, manager accountability practices, embedding diversity goals into organizational strategy (Dobbin and Kalev, 2015; Richard et al., 2013), as well as initiatives such as bystander intervention and ally networks. These would be better captured through a review of the primary literature.

Contrary to earlier reviews (Colella et al., 2017; Triana et al., 2021), which raised concerns over an over-representation of research on race and ethnicity diversity, and an under-emphasis on disability, our findings revealed a paradox in the opposite direction. That is, disability (and to a lesser extent, age) dimensions of diversity were well represented among the included articles, and along with them, interventions such as workplace accommodations and health promotion activities.

This is unlikely to be due to the search terms, as we intentionally avoided searching based on diversity dimensions (e.g. “race”) to prevent biasing results towards one dimension. One possibility is that given that D&I policies had their starting point in race diversity, research on interventions for race diversity are the “default” and are studied less frequently in isolation to the exclusion of other diversity dimensions. This suggests that while the interventions for race diversity may generalize to other dimensions (e.g. diversity training, de-biasing selection procedures), the reverse may not be true and the findings may be less useful for organizations interested in only racial D&I.

One explanation for the over-representation of the disability dimension that may be related to our search strategy is that we purposefully targeted high-quality empirical reviews, including systematic reviews and meta-analyses. The results may reflect the greater use of these methods (along with intervention-focused research) for studying disability, versus other dimensions that tend to receive more attention in conceptual and theoretical papers. To illustrate, 12 of the 14 included articles mentioning disability in their title were systematic or scoping reviews, and 9 of these came from healthcare and rehabilitation journals—disciplines where systematic reviews are common given their origins in clinical medicine (Chalmers et al., 2002).

In their landmark study on the efficacy of corporate diversity policies, Kalev et al. (2006) cautioned against so-called best practices, noting that “[w]e know a lot about the disease of workplace inequality, but not much about the cure” (p. 590). This meta-review brings together intervention-focused research across multiple knowledge bases to answer “what works”. It is intended to serve as key reference to assess and compare the quality of evidence and empirical outcomes for a catalogue of D&I interventions. The results help bridge the research–practice gap described by Kulik (2014). We have provided evidence on the policies and programs components of diversity management systems and offer a decision-making resource and clear expectations for practitioners aiming to understand what interventions can be carried out for which diversity dimensions, and what outcomes to expect when implementing them. Although there is still much detail to fill in, we have provided a starting point in making the vast literature of D&I more accessible, relevant, and useful for practitioners.

Funding for this work was provided in part by Inclusive Australia (www.inclusiveaustralia.com.au), a not-for-profit organization seeking to promote change for social inclusion in Australia. Liam Smith is on the board of Inclusive Australia.

The Supplementary Material for this article can be found online.